bredelings / bali-phy

Bayesian co-estimation of phylogenies and multiple alignments via MCMC

Install

If you just want to install bali-phy, please visit the release page. If you want to compile BAli-phy from source, the quick start instructions are below.

Compiling

You will need a C++ compiler that understands C++20.

• gcc 11 (or higher) works
• clang 14 (or higher) works
• XCode 14 (or higher) works

Install Prerequisites

On Ubuntu 24.04 (or higher), or on Debian testing, you can use apt-get:

sudo apt-get install g++ libcairo2-dev meson libboost-all-dev

On Mac (or Linux, actually) you can use homebrew:

brew install cairo meson boost pkg-config

On miniconda, you can use:

conda create -n devel -c conda-forge --strict-channel-priority
conda activate devel
conda install meson gxx boost-cpp cmake pkg-config cairo
export BOOST_ROOT=$CONDA_PREFIX

Meson version

After installing prerequisites above, check that the meson version is at least 1.1.

meson --version

If your version of meson is less than 1.1, you need to install the most recent version of meson into a python virtual environment:

python3 -m venv meson_venv
source meson_venv/bin/activate
pip install meson ninja

Build BAli-Phy

This will build the latest unreleased beta version of BAli-Phy, which fixes some memory issues in 3.6. There are some changes to the model language. Check the NEWS file for the details.

git clone https://github.com/bredelings/BAli-Phy.git
cd BAli-Phy
meson setup build --prefix=$HOME/Applications/bali-phy
ninja -C build install
ninja -C build test

Adding bali-phy to your $PATH

In order to run the installed software, you should add bali-phy to your $PATH.

Installed locations

If you installed in $HOME/Applications/bali-phy/ as recommended above, then files will be in:

Location	Files
~/Applications/bali-phy/bin	Binary executables.
~/Applications/bali-phy/share/bali-phy/examples/sequences	Example files.
~/Applications/bali-phy/share/doc/bali-phy/	Documentation.

Further Documentation

• http://bali-phy.org/
• Manual
• Tutorial

The Manual describes how to install bali-phy in greater detail.

bali-phy's Issues

bp-analyze (logging.debug)

Hello - is this known?
BAli-Phy running fine on test with [--iter=150] and finishes, but post-analysis with bp-analyze [../bp-analyze dir-1 dir-2] produces this:

~/../bali-phy-3.5.0/bin/bp-analyze", line 405 logging.debug(f"LogFileRun: {mcmc_output}") ^

System is BioLinux 8 Ubuntu 14.04 LTS. BAli-Phy version is latest (3.5, BUILD: Apr 20 2020 14:13:27, ARCH: linux x86_64)
Same result appears with no arguments fed in to bp-analyze. Reasonable fiddling doesn't help..
Any comment much appreciated.

Leo

gcc 10.1 compilation fails, call of overloaded ptree is ambiguous

gcc 10 has some improved pgo and lto options that might offer better performance.

However, compilation fails for me, with the message:
./BAli-Phy-3.5.0.1/src/models/rules.cc:362:51: error: call of overloaded ‘ptree()’ is ambiguous
...
./BAli-Phy-3.5.0.1/src/models/rules.cc:374:62: error: call of overloaded ‘ptree()’ is ambiguous

Compilation worked for me under gcc 9.3.0.

Build output attached.

build.log

Add alignment constraints to handle long sequences

When aligning long sequences we need to constrain the alignment in order to avoid taking too much CPU time (and RAM).

• allow constraints that residue x[i] is within some distance of residue y[j] instead of requiring that they align exactly.
• allow alignments within some "envelope" around an initial alignment

The problem here is that dynamic programming for 3 or more sequences allows constraints like x[i] <= y[i]. However, we must then decompose the alignment of three sequences into three pairwise alignments. This relaxes the ordering imposed encoding the alignments into paths through the dynamic programming matrix.

Ideas

• Can we impose "not-less-than" constraints instead of "<=" constraints?

Builds are not available from website

The current (stable) releases are not available from the website (e.g. http://bali-phy.org/files/bali-phy-2.3.8-linux64.tar.gz but also the versions for OSX and Windows).

Please check the free() or other scripts related with memory

OS: ArchLinux
Installation: https://github.com/BioArchLinux/Packages/blob/master/BioArchLinux/bali-phy/PKGBUILD

When I finished running this problem and then It is still running in the background and quickly fills up the CPU. I don't want to repeat it, but it's still a problem for other users.

Test suite error in 4.0-beta7

Hi,
I tried to package 4.0-beta7 for Debian. In our CI you can find a full build log which contains an error in one test:

5/37 bali-phy 5d -A 200             FAIL             10.83s   killed by signal 11 SIGSEGV
>>> ASAN_OPTIONS=halt_on_error=1:abort_on_error=1:print_summary=1 MALLOC_PERTURB_=129 UBSAN_OPTIONS=halt_on_error=1:abort_on_error=1:print_summary=1:print_stacktrace=1 /builds/med-team/bali-phy/debian/output/source_dir/obj-x86_64-linux-gnu/src/bali-phy/bali-phy /builds/med-team/bali-phy/debian/output/source_dir/obj-x86_64-linux-gnu/../examples/5S-rRNA/5d.fasta --iter=200 --package-path=/builds/med-team/bali-phy/debian/output/source_dir/obj-x86_64-linux-gnu/src/builtins:/builds/med-team/bali-phy/debian/output/source_dir -Inone

Please let me know if you need further information.
Kind regards, Andreas.

segmentation fault

For one of my input files, with release 4.0-beta7, I get:

Allocation failed in sample_tri_multi!  Proceeding.
Allocation failed in sample_tri_multi!  Proceeding.
Allocation failed in sample_tri_multi!  Proceeding.
[maybe 30 times]
[1]    25144 segmentation fault  bali-phy APOB_extracted_oneseq_onedirection_All.fasta -S gtr -n  -i 2000

I then tried building from source with the current master branch to see if this was something that had gotten resolved, and that gave me:

bali-phy: Error! evaluating reg # 8295 (unchangeable): case <175495> of {_ -> <175496>}

evaluating reg # 175495 (unchangeable): case <175496> of {_ -> <2404>}

evaluating reg # 175496 (unchangeable): MCMC:runMCMC 2000 0 0

evaluating reg # 471014 (unchangeable): case <86522> of {_ -> <86561>}

evaluating reg # 86522 (unchangeable): case <86561> of {_ -> <2404>}

evaluating reg # 86561 (unchangeable): MCMC:walk_tree_sample_alignments <6084> <46054> 0 0

std::bad_array_new_length

I can provide the input fasta file if that would be helpful. Any thoughts?

Thanks!

<bitset> needs inclusion in src/dp/hmm.cc

BAli-Phy 3.6.1
gcc 12.2.1
glibc 2.36

Compilation error, bitset undefined. Fixed by this:

`diff --git a/src/dp/hmm.cc b/src/dp/hmm.cc
index 8be51b5..ea8a15e 100644
--- a/src/dp/hmm.cc
+++ b/src/dp/hmm.cc
@@ -24,6 +24,7 @@
///

#include
+#include
#include "hmm.H"
#include "2way.H"
#include "imodel/imodel.H"
`

BAli-Phy on conda?

Hi There! I was curious if there is any ongoing efforts to include BAli-Phy as a package available through bioconda?

use eigen instead of ublas/tnt/jama

I considered doing this, so I looked at the file
https://github.com/bredelings/BAli-Phy/blob/master/src/math/exponential.C
but I got scared away because its scope includes only time-reversible rate (or transition) matrices, and because it has a weird gamma thing that I don't understand and probably don't want to use, and because the testing is weird so I won't know if I've screwed up the project. The test is controlled by an #ifdef?

install prerequisites should mention boost

This is a fairly major dependency.

Ubuntu LTS?

It is nice that BAli-Phy is available using package managers. However, it is not available on the current Ubuntu LTS release (v18.04).

Is it possible to make it available from an LTS (which I imagine most users, er, use), or are there significant roadblocks?

If this is not possible, it is not a big deal, as building it is straightforward (and the precompiled binaries are almost as convenient).